‘Attribution studies help increase awareness about links between human activities and climate change’
Libya saw devastating floods in September 2023. Photo: @UNDP / X (previously Twitter)
In October 2023, scientists from the Chinese Academy of Sciences and the Met Office Hadley Centre, United Kingdom released a paper on the attribution of extreme heat and drought events in 2022. It found that the extreme drought covered 47.3 per cent of global land areas in 2022, which was the second most widespread year since 1980. The group is in the process of conducting a similar analysis for extreme events in 2023.
Down To Earth caught up with Wenxia Zhang from the Chinese Academy of Sciences and one of the authors of the paper at the 28th Conference of Parties (COP28) to the United Nations Framework Convention on Climate Change to speak about why extreme weather attribution studies are important, what improvements it has seen in the recent past and how such studies feed into COP.
Edited excerpts:
Rohini Krishnamurthy: You have released papers on the overview of extreme weather and climate events in 2021, and 2022, and are in the process of preparing one for 2023. What are some trends that have emerged in your analysis?
Wenxia Zhang: There is a significant increase in the trend of heatwaves almost everywhere globally. This is summarised in the latest Intergovernmental Panel on Climate Change’s (IPCC) sixth assessment report (AR6).
In terms of heavy rainfall and droughts, the trend is not as significant as extreme temperatures because the processes governing extreme rainfall and droughts are much more complicated. Also, observations of rainfall extremes are limited in some regions.
In fact, in attribution studies, we estimate whether and to what extent human activities have changed the intensity or frequency of these extreme events. Along with attribution, we also assess how confident or how certain these conclusions are.
For extreme temperatures, we have very high confidence in the attribution results. This means we have high confidence that human activities have increased heatwaves. But for extreme rainfall or tropical / extra-tropical cyclones or droughts, our confidence in attributing their change to human influences is only low to medium.
I am also interested in rainfall induced by cyclones. In the IPCC AR6 assessment report, there are some conclusions about the change in cyclones and the general conclusion is that as the climate warms, the frequency of tropical cyclones does not show a significant change, but there will be more very intense cyclones.
So that means when the cyclone occurs, it tends to become stronger and the related rainfall will also be enhanced. This is the understanding of the long-term trend and I think in these years we seem to have witnessed such an emerging trend. The Libyan and North China flooding in 2023, both related to hurricanes or typhoons, are a good example of this.
RK: Why are the processes much more complicated with rainfall and wildfire events, for example?
WZ: Climate or extreme events are affected by both external forcings and internal climate variability. The external forcings usually result in a long-term trend (forced climate change signal), such as an increasing trend in heatwaves. Climate internal variability causes fluctuations at different time scales, such as inter-annual or decadal scales, in response to the forced climate change signal.
In attribution studies, we need to separate the externally forced signals from the internal variabilities. For rainfall events, droughts or wildfires, the internal variabilities are much larger (than for temperature extremes). That means it’s harder for forced climate change signals to emerge from these very large natural fluctuations.
Generally, the detectability would increase if we had data from a longer period. For example, for extreme rainfall, we have daily rainfall observations since the 1950s, which is just 70 years of data. This is not long enough to allow human signals to emerge from climate noise in some regions.
RK: The World Weather Attribution (WWA) also comes up with attribution studies. How are these studies different from formal ones?
WZ: The WWA consists of a group of scientists making huge contributions to rapid attribution by providing conclusions usually about weeks after an event has occurred. This information is very important to the public and policymakers to have a very first understanding of how these events occurred and to what extent they are affected by climate change or human activities.
These rapid attribution conclusions also provide valuable information to decision-making in terms of adaptation and recovery.
Because rapid attribution is desired, WWA uses a relatively simple method. It compares the occurrence probability in the later period and an earlier period. It does not involve physical processes.
The most widely used attribution method uses simulations with and without human influence. By comparing the two, we can understand how humans have affected these events.
A more formal attribution provides a more detailed and comprehensive conclusion with a detailed physical understanding of these events. Different methods of attribution have been developed in recent years.
One example is the storyline approach. Using this, scientists first identify the key physical processes behind this event. For example, for an extreme rainfall event related to tropical cyclones, we know that the key circulation configuration is a tropical cyclone. Then, we conduct a pair of climate model simulations by prescribing the circulations to the observed real circulation.
And we change the thermodynamic background in the simulations. For example, we give a thermodynamic background in the present day in one set of simulations, and in the other set, we give the thermodynamic background in an earlier period (in the 1950s, for example). By comparing the rainfall events in these two sets of simulations, we can understand how rainfall changes with human influences over the past decades if the same circulation occurs in both periods.
It is a good practice to use different methods of attribution to assess different aspects of extreme events, which then indicates the confidence level of the attribution results. For example, the probability-based attribution approach tells you to what extent humans have changed the probability of this event and the storyline approach tells you the extent to which human activities have intensified or weakened the extreme event given the same circulations. From these different aspects, we can form a more comprehensive conclusion.
RK: How have attribution studies changed in the recent past?
WZ: Attribution studies have developed very fast in the past two decades. The first improvement is the development of more methods like the storyline approach.
Another improvement is that there are more and more attribution simulations with fine experiment designs to explicitly address the influences of different forcing agents. In addition, the climate models used for attribution have finer spatial resolution in recent years. With higher resolution, we can better depict the extreme events.
The third improvement is the addition of a more physical understanding of attributions. For example, in an attribution study, we looked at how human activities have affected these two different aspects of the heavy rainfall events in East Asia: Daily intensity (maximum rainfall falling in just one day) and the accumulated rainfall total over a persistent period (over a month, for example). We find that human activities have intensified the daily rainfall.
But for the persistent accumulated rainfall, we find that human activities over the past decades have reduced the heavy rainfall amount that persists on a monthly scale.
This is because different mechanisms are playing a dominant role. For the daily intensity, the increased intensity is dominated by greenhouse gas forcing via increasing atmospheric moisture, which then fuels heavy rainfall.
But on a monthly scale, persistent heavy rainfall and aerosols (fine solid or liquid particles suspended in the atmosphere) released by human activities dominate. Aerosols have weakened the persistent heavy rainfall in East Asia over the past decades. This is associated with the aerosol-induced weakening of large-scale East Asian summer monsoon circulation, which serves as a dynamic background for persistent rainfall events.
By adding more physical understanding like this, we understand why human activities increase the daily rainfall but decrease the persistent heavy rainfall in East Asia. With such physical understanding, we have higher confidence about the attribution.
RK: How do attribution studies feed discussions at COP?
WZ: Firstly, attribution studies help increase awareness about the linkage between human activities and climate change or specific events. This enhanced understanding fosters a greater sense of urgency for policymakers, the public, and industries to take action to reduce carbon emissions.
Second, it can support decision-making and adaptation efforts. When policymakers better understand the extent of human influence on different types of extreme events, they can take more targeted and efficient policies for mitigation and adaptation.
We also want to pay more attention to the climate change problem in the Global South, which is vulnerable to climate change and has suffered a lot from climate hazards in recent decades. Fine extreme climate prediction products at high resolution could help countries develop more efficient adaptation strategies.
But these require very heavy computing resources. Developing countries may not have such resources, so global efforts and international collaborations are highly valued in this regard, to help developing nations prepare themselves for extreme events. As COP28 has highlighted, we need to be “uniting for climate action”.